There’s a compliment production people love to receive: “I don’t know what we’d do without you.” It sounds like job security. It’s a defect report.
If a process falls over when you step away, you didn’t build a process; you became one. The team doesn’t have a system; it has a person with good habits. And people leave, get promoted, get sick, take vacations. Every save that needs you personally is evidence that the system still has a hole shaped exactly like you.
The strongest evidence I can offer for any system I’ve built is who ran it after me.
The crash-triage loop I describe in The dashboard is not the system took a senior producer to stand up: a daily cadence, a studio-visible status page, a named owner for every crash class, callstacks searchable in the tracker. Building it meant earning buy-in across QA, production, and engineering by influence rather than rank, and writing down state that had only ever lived in people’s heads. Running it, once it existed, was a different job. By the time I left, a production coordinator was running it day to day.
And the handoff wasn’t an accident of my departure. Once a system proves its worth, I start delegating it: hand the daily loop to someone who can run it, stay close enough to catch a wobble, then take the attention it frees to the next problem the team needs solved. That’s what building it well buys. The team keeps the system, and I go fix the next thing.
That drop in required seniority is the measure. Not the target the system hit. Whether it still hits the target when the person who built it steps away.
What survives a handoff tends to be boring on purpose. Truth written on a page, so nobody has to remember it. A named owner on every problem, because “someone should watch this” means nobody is watching it. History you can search, so the next person doesn’t need to have been in the room. And a cadence simple enough to keep on a bad week. A daily loop with one question beats a brilliant workflow that dies by week two.
It’s the same reason I self-host my own infrastructure. Running your own services is a standing exercise in the discipline: they stay up because they’re understood, documented, and monitored, not because I remember to go check on them. When something breaks at 2 a.m., the fix has to be findable by a version of me who’s tired and has forgotten the details, which is functionally a different person. Build for your replacement and you build better systems, even when the replacement is just future you.
This site makes a promise about permanence: fixed for good, meaning both permanent and correct. Permanence has to include the day you’re not there.
Build so you can leave.